This research examines differential adult mortality in the US according to multiple causes of death. Specifically, the research: (1) constructs measures of multiple cause of death, (2) estimates multiple cause specific mortality differentials by demographic and social characteristics, and (3) compares these differentials with analogous estimates based on underlying cause of death. The research seeks to answer the question of whether the greater complexity of multiple cause of death measures provides sufficient benefit to researchers in terms of giving substantively different findings or conclusions in regard to adult mortality differentials. Data come from the National Health Interview Survey - Multiple Cause of Death (NHIS-MCD) linked file for 1986-1995, with updates through 1997 as soon as these are released by the National Center for Health Statistics. These data combine survey responses on demographic, social, and health characteristics for over 600,000 adults in the civilian non- institutionalized population and included in the NHIS in survey years 1986-1994 with follow-up information from the National Death Index on 38,849 deaths through 1995. Preliminary indications are for a total of 54,534 deaths through 1997. The research begins with and then extends an extant scheme of five basic underlying causes (circulatory, cancer, respiratory, social pathologies, and residual) to provide more detail for important medical conditions (e.g., considering heart disease and stroke separately, as well as looking specifically at diabetes, Alzheimer's Disease, anal other conditions). Multiple causes are measured as the combination of these categories based on the inspection of all medical conditions listed on the death certificate, without regard to whether a particular condition is listed as underlying. The research examines age- adjusted mortality differentials by sex, race/ethnicity, education, family income, occupational status, and self-reported health, using proportional hazards modeling with multiple outcomes reflecting the various causes of death or survival. Statistical adjustments are included for design features of the NHIS in order to produce accurate parameter estimates and unbiased standard errors and confidence intervals.